Transform coding of audio

Original audio: Mono, 16 bits/sample, 44100 sample/s, 705.6 kbps

Zonal coding. Lloyd-Max quantization. 2 bits/sample (88.2 kbit/s)

• No transform. SNR 9.78 dB.
• DCT, block size 2. SNR 17.96 dB.
  Bit allocation: 4 0
• DCT, block size 4. SNR 19.29 dB.
  Bit allocation: 5 2 1 0
• DCT, block size 8. SNR 21.67 dB.
  Bit allocation: 6 3 2 2 2 1 0 0
• DCT, block size 64. SNR 23.65 dB.
  Bit allocation: 8 6 4 4 3 3 3 3 3 4 4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 2 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Uniform quantization (same stepsize for all transform components). Huffman coding. 2 bits/sample (88.2 kbit/s).

• No transform. SNR 9.46 dB.
• DCT, block size 2. SNR 15.19 dB.
• DCT, block size 4. SNR 20.86 dB.
• DCT, block size 8. SNR 24.81 dB.
• DCT, block size 64. SNR 29.66 dB.

JPEG coding of still image

Original image: 768x512 pixels, 24 bits/pixel (RGB)

JPEG, 25 times compression, 0.96 bits/pixel

JPEG, 50 times compression, 0.48 bits/pixel

JPEG, 100 times compression, 0.24 bits/pixel

JPEG, 200 times compression, 0.12 bits/pixel

JPEG, 400 times compression, 0.06 bits/pixel